Epigenetic Regulation of Replication-Dependent Histone mRNA 3 End Processing

(1)

Epigenetic Regulation of Replication-Dependent Histone mRNA 3’ End Processing

Dissertation

for the award of the degree

“Doctor rerum naturalium (Dr. rer. nat.)”

Division of Mathematics and Natural Sciences of the Georg-August-Universität Göttingen

submitted by Judith Pirngruber

born in

Oldenburg, Germany

Göttingen, 2010

(2)

Prof. Dr. Steven A. Johnsen

Doctoral Committee:

Prof. Dr. Steven A. Johnsen (1^st Referee)

Molecular Oncology,

University of Göttingen Medical School, Göttingen

Prof. Dr. Heidi Hahn (2^nd Referee)

Molecular Developmental Genetics,

Prof. Dr. Holger Reichardt

Cellular and Molecular Immunology,

Date of the oral examination: 28^th April 2010

(3)

I hereby declare that the PhD thesis entitled “Epigenetic Regulation of Replication- Dependent Histone mRNA 3’ End Processing” has been written independently and with no other sources and aids than quoted.

____________________________________

Judith Pirngruber

March, 2010

Göttingen, Germany

(4)

Für Opa Bruno

„Opa, ich denk’ an Dich, wenn ich das Gänseliesel hochkletter’.“

(5)

List of Abbreviations

(NH₄)₂SO₄ ammonium sulphate

36B4 mouse homolog of human RPLP0 gene

5-FU 5-Fluorouracil

7-AAD 7-amino-actinomycin D

7SK snRNA 7SK small nuclear RNA

ActD Actinomycin D

ADP adenosine diphosphate

APC/C anaphase-promoting complex/cyclosome

Asf1 anti-silencing factor 1

ASH2L absent, small, or homeotic-like protein

ATM ataxia telangiectasia mutated

ATP adenosine triphosphate

BGP beta-glycerophosphate

bp base pair

BRD4 bromodomain containing 4 protein

BrdU 5-bromodeoxyuridine

Bre1p yeast brefeldin A sensitive protein 1

Bur1, 2 bypass upstream activating sequence requirement protein 1, 2

CAK cyclin-dependent kinase activating kinase

CB cajal body

CBC cap binding complex

CBP80 cap binding protein 80

CCND1 cyclin D1

CDC cell division cycle

CDK cyclin-dependent kinase

cDNA complementary DNA

CE capping enzyme

CF cleavage factor

CHD1 chromodomain protein 1

ChIP cromatin immunoprecipitation

Cip cyclin inhibitor protein

CKI cyclin-dependent kinase inhibitor

COMPASS complex of proteins associated with Set1

Cont. control

CPSF cleavage and polyadenylation specifity factor

CSTF cleavage stimulating factor

CTD carboxy-terminal domain

Ctr9p yeast cyclin three requiring 9 protein

CUL1 cullin 1

DCP1, 2 decapping protein 1, 2

DHFR dihydrofolate reductase

DISC death-inducing signaling complex

dMes-4 Drosophila maternal effect sterile family member 4

DMSO dimethyl sulfoxide

DNA deoxyribonuleic acid

DOT1L disruptor of telomeric silencing 1-like protein

DP1, 2 dimerization partner 1, 2

DRB 5,6-dichloro-1-β-D-ribofuranosyl-benzimidazole

DSE downstream sequence element

DSIF DRB sensitivity-inducing factor

E(z) enhancer of zeste

E1 ubiquitin activating enzyme

E2 ubiquitin conjugating enzyme

E2F E2 promoter binding factor

E3 ubiquitin-protein-isopeptide ligase

EDTA ethylenediaminetetraacetic acid

EIF4G eukaryotic translation initiation factor 4-γ

ERα estrogen receptor α

(8)

EST expressed sequence tags

et al. et alii (and others)

F forward

f. and the following page

FACT facilitates chromatin transcription

FBS fetal bovine serum

FCP1 TFIIF-associated CTD phosphatase 1

FCS fetal calf serum

ff. and the following pages

Fig Figure

FIP1 factor interacting with poly(A) polymerase 1 protein

FLASH FADD-like IL-1β-converting enzyme associated huge protein

FLICE FADD-like IL-1β-converting enzyme

FP flavopiridol

g relative centrifugal force

gal1 yeast galactokinase 1 gene

GAPDH glyceraldehyde-3-phosphate dehydrogenase

GST glutathione S-transferase

GTF general transcription factor

h hour

H1, H2A, H2B, H3, H4 histone 1, 2A, 2B, 3, 4

H1299 human lung cancer cell line

H2Bub1 histone H2B monoubiquitination

H3K27me3 histone H3 lysine 27 trimethylation H3K36me2, 3 histone H3 lysine 36 di-, trimethylation H3K4me1, 2, 3 histone H3 lysine 4 mono-, di-, trimethylation H3K79me2, 3 histone H3 lysine 79 di-, trimethylation

HAT histone acetyltransferase

HBF hairpin binding factor

HCF-1 host cell factor-1

HCT116 human colon carcinoma cell line

HDAC histone deacetylase

HDE histone downstream element

HEK293 human embryonic kidney cell line

HeLa human epithelial cervical cancer cell line (from Henrietta Lacks) HEXIM1, 2 hexamethylene bisacetamide-induced protein 1, 2

hHR6 human homolog of Rad6p

HINFP histone H4 transcription factor

HIST1, 2, 3 histone cluster 1, 2, 3

HIV1 human immunodeficiency virus 1 gene

HLB histone locus body

HOX homeobox

HPV human papilloma virus

hsp70 heat shock protein 70 gene

HU hydroxyurea

HYPB huntingtin interacting protein b

i.e. id est (that is)

IgG immunoglobulin G

ILC invasive lobular carcinoma

INK inhibitor of cyclin-dependent kinase

IP immunoprecipitation

K120, 123 lysine residue 120, 123

kb kilobase pair

kDa kilodalton

Kin28p yeast protein kinase homolog to CDC28

Kip kinase inhibitor protein

KM KM05283

KMT lysine methylating transferase

LARP7 La ribonucleoprotein domain family, member 7

Leo1p yeast left open reading frame on plasmids pK9 and pUHC15-2 protein

LSM sm-like protein

(9)

M molar

MDM2 murine double minute 2

MgCl₂ magnesium chloride

min minute

ml milliliter

MLL mixed lineage leukemia

mM millimolar

M-MuLV Moloney murine leukemia virus

mRNA messenger RNA

MTRNR2 mitochondrially encoded 16S RNA

MyoD myogenic differentiation protein

n number of individual values

NaCl sodium chloride

NAP1 nucleosome assembly protein 1

NCBP1 nuclear cap binding protein subunit 1

Neg. negative

NELF negative elongation factor

NEM N-Ethylmaleimide

NIMA never in mitosis A protein

NP-40 nonyl phenoxylpolyethoxylethanol

NPAT nuclear protein Ataxia Telangiectasia locus

NSD1 nuclear receptor-binding SET domain-containing protein 1

nt nucleotide

NTP nucleotide triphosphate

Nutlin Nutley inhibitor

OCA-S OCT1 coactivator in S phase

OCT1 octamer-binding protein 1

ORF open reading frame

P53-/-, p21-/- p53 null cell line, p21 null cell line

P53+/+ p53 wild type cell line

PABP poly(A)-binding protein

PAF polymerase associated factor

PAP poly(A) polymerase

PBS Phosphate Buffered Saline

PCF11 protein 1 of cleavage and polyadenylation factor I PCNA proliferating cell nuclear antigen

PCR Polymerase Chain Reaction

PE phycoerythrin

PHD plant homeodomain

Phe phenylalanine

PI propidium iodide

PIC preinitiation complex

PIC Protease Inhibitor Cocktail

Pin1p yeast peptidyl-prolyl cis/trans isomerase, NIMA-interacting 1 protein PIP7S P-TEFb interaction protein for 7SK stability

PITALRE Pro-Ile-Thr-Ala-Leu-Arg-Glu motif protein Pob3p yeast polymerase I binding protein 3

PPARγ peroxisome proliferator-activated receptor gamma

PRC2 polycomb repressive complex 2

Pro proline

P-Ser phosphorylated serine residue

P-TEFb positive transcription elongation factor b

PTM posttranslational modification

qRT PCR quantitative reverse transcriptase PCR

R reverse

Rad6p yeast radiation-sensitive mutant protein 6

RAP80 receptor-associated protein 80

RB retinoblastoma protein

RBBP5 retinoblastoma binding protein 5

RBP1 largest subunit of RNAPII

RNA ribonuleic acid

(10)

RNAPII RNA Polymerase II

RNAPII₀ hyperphosphorylated form of RNA Polymerase II RNAPII_a hypophosphorylated form of RNA Polymerase II

RNF ring finger protein

RPLP0 ribosomal protein, large, P0

Rtf1p yeast restores TBP function protein 1

Rtt103p yeast regulator of Ty1 transposition 103

s.d. standard deviation

SAGA Spt-Ada-Gcn5-acetyltransferase

SCF SKP1-CUL1-F-box-protein complex

SCP small CTD phosphatase

Sdi1 senescent cell-derived inhibitor 1

Ser serine

Set1p, Set2p yeast (Su(var)3-9, E(z) and Trithorax)-containing protein 1, 2

SKP1 S phase kinase-associated protein 1

SLBP stem-loop binding protein

SLIP1 SLBP-interacting protein 1

SmD1 small nuclear ribonucleoprotein D1 polypeptide 16kDa SmD2 small nuclear ribonucleoprotein D2 polypeptide 16.5kDa

SMYD2 MYND domain containing 2

snRNP small nuclear ribonucleoprotein

Spt suppressor of Ty

SSBP subtype specific binding protein

SSRE subtype specific regulatory element

SSRP1 structure specific recognition protein 1

Su(var) suppressor of variegation

SUPT4H, SUPT5H, SUPT16H human suppressor of Ty4, 5, 16 homolog

Taq Thermus aquaticus

TBP TATA-binding protein

TEC transcription elongation complex

TFF1 trefoil factor-1

Thr threonine

TIP60 Tat interacting protein, 60kDa

TP53BP1 tumor protein p53 binding protein 1

TR transcribed region

Tris-HCl Tris(hydroxymethyl)aminomethane hydrochloride

Trp tryptophan

TRRAP transformation/transcativation domain-associated protein

TSS transcription start site

TUTase terminal uridyltransferase

Tyr tyrosine

U unit

U2 OS human osteosarcoma cell line

U2AF U2 small nuclear RNA auxiliary factor

UBCH6 ubiquitin-conjugating enzyme 6

UBE2A, B ubiquitin-conjugating enzyme E2A, B (Rad6p homolog)

UBP ubiquitin protease

UIM ubiquitin interaction motif

UPF1 up-frameshift protein 1

UPS ubiquitin-proteasome system

USP22 ubiquitin specific protease 22

v/v volume per volume

w/v weight per volume

Waf1 wild type p53-activated fragment 1

WDR5 WD repeat domain 5 protein

WEE1 “small” mutant protein 1

WT wild type

ZFP100 zinc finger protein 100

µl microliter

µm micrometer

(11)

List of Figures

General Introduction, General Discussion, Summary and Conlusions

Fig 1 RNAPII-dependent transcription and the CTD phosphorylation... 6

Fig 2 Posttranslational modifications of the core histones and their positions ... 11

Fig 3 The mammalian cell cycle... 19

Fig 4 Formation of the processing complex on canonical histone pre-mRNAs... 26

Fig 5 Model for the function of NPAT in activating histone genes ... 28

Fig 6 The p53/RB/E2F pathway ... 31

Fig 7 Schematic diagram of NPAT... 106

Fig 8 Schematic representation for the role of G₁/G₀ arrest in controlling the mode of replication-dependent histone 3’ end formation ... 116f. Publication I Fig I.1 H2Bub1 depends on CDK9 activity rather than transcriptional activation per se . 37 Fig I.2 CDK9 activity is essential for maintaining global levels of H2Bub1... 38

Fig I.3 CDK9 and H2Bub1 direct replication-dependent histone messenger RNA 3’-end processing but not transcription ... 41

Fig I.4 CDK9 activity recruits proteins involved in H2B ubiquitination and histone messenger RNA 3’-end formation, and decreases RNAPII read-through ... 42

Supp Fig I.S1 CDK9, RNF20 and RNF40 are essential for maintaining global levels of H2B monoubiquitination... 51

Supp Fig I.S2 CDK9, RNF20/40 and associated proteins and chromatin modifying enzymes specifically affect replication-dependent histone mRNA polyadenylation ... 52

Publication II Fig II.1 CDK9 activity is essential for H3K4me3 and H3K36me3 ... 56

Fig II.1 CDK9 and RNF20/40 knockdown increase the formation of spliced HIST1H2BD (A) and HIST1H2AC (B) transcripts ... 60

Fig II.2 Schematic representation of the role of CDK9 in controlling replication-dependent histone mRNA 3' end processing ... 65

Publication III Fig III.1 P53 accumulation reciprocally regulates replication-dependent histone gene expression and mRNA polyadenylation in a p21-dependent manner ... 70

Fig III.2 Accumulation of p53 decreases the recruitment of CDK9, PAF1 and NPAT... 72

Fig III.3 Accumulation of p53 and induction of p21 precede cell cycle arrest, decreases in replication-dependent histone and NPAT gene expression and ultimately accumulation of polyadenylated histone transcripts ... 75

Fig III.4 Accumulation of polyadenylated replication-dependent histone mRNAs is induced following a G₁/G₀ cell-cycle arrest... 77

Fig III.5 NPAT expression is essential for optimal levels of replication-dependent gene transcription and proper pre-mRNA 3’ end processing ... 79

Fig III.6 NPAT knockdown mimics p53-induced G1 cell-cycle arrest and is required for optimal recruitment of CDK9 and PAF1... 81

Supp Fig III.S1 Nutlin-3a treatment does not affect cell cycle progression in the absence of p53 or p21 ... 87

Supp Fig III.S2 Nutlin-3a treatment does not significantly induce apoptosis in HCT116 cells irrespective of the p53 and p21 status ... 88

Supp Fig III.S3 Total H2B protein levels do not change following Nutlin-3a treatment... 88

Supp Fig III.S4 Nutlin-3a time-dependently decreases the fraction of HCT116 cells in S phase... 89

Supp Fig III.S5 NPAT downregulation depends upon an intact p53-p21 axis ... 90

(12)

Supp Fig III.S6 Accumulation of polyadenylated replication-dependent

histone mRNAs is induced following a G₁/G₀ cell-cycle arrest ... 91 Supp Fig III.S7 The effect of Cyclin E1 on replication-dependent histone gene expression

and pre-mRNA 3’ end processing is independent of p53 status ... 92 Supp Fig III.S8 NPAT expression is essential for optimal levels of replication-dependent

gene transcription and proper mRNA 3’ end processing... 93 Supp Fig III.S9 Decreased CDK9, PAF1 and NPAT recruitment to replication-dependent

histone genes is a specific effect of NPAT knockdown and not due to

non-specific effects of siRNA transfection... 94

(13)

List of Tables

Publication I

Supp Table I.S1 Antibodies used for ChIP and western blot analyses and the respective dilutions 48 Supp Table I.S2 Primers utilized in ChIP in 5’ to 3’ orientation ... 49 Supp Table I.S3 Primers utilized in RT-PCR in 5’ to 3’ orientation... 49 Supp Table I.S4 siRNAs utilized in a 5’ to 3’ orientation... 50

Publication III

Supp Table III.S1 siRNAs utilized for knockdown studies in a 5’ to 3’ orientation... 85 Supp Table III.S2 Antibodies used for ChIP, IP, western blot analyses and

Immunofluorescence and the respective dilutions... 85f.

Supp Table III.S3 Primers utilized for chromatin immunoprecipitation

analyses in 5’ to 3’ orientation ... 86 Supp Table III.S4 Primers utilized for gene expression analyses in 5’ to 3’ orientation ... 86f.

(14)

Abstract

Cyclin-dependent kinase 9 (CDK9) was shown to be the Ser/Thr kinase that phosphorylates the C-terminal domain (CTD) of the largest subunit of RNA Polymerase II (RNAPII) at serine residue 2 in mammalian thus regulating its activity following transcriptional inititiation. Furthermore, CDK9 was suggested to also regulate co- transcriptional histone modifications and mRNA processing events. Our studies show that CDK9 functions in maintaining global levels of histone H2B lysine 120 monoubiquitination (H2Bub1) at lysine 120 and guides a complex network for additional histone modifications, including histone H3 lysine 4 trimethylation (H3K4me3) and H3K36me3. However, these modifications seem to not only be dependent upon phosphorylation of the CTD but include other CDK9 targets like suppressor of Ty 5 homolog (SUPT5H), negative elongation factor-E (NELF-E) and probably the ubiquitin-conjugating enzyme E2A (UBE2A). Interestingly, we found that CDK9 activity and the mentioned histone modifications are necessary for maintaining proper replication-dependent histone mRNA 3’ end processing since CDK9 knockdown resulted in an inefficient recognition of the correct cleavage site and led to read-through of RNAPII to an alternative downstream polyadenylation signal. Furthermore, induction of a G₁ cell cycle arrest by accumulation of p53 resulted in an increase in polyadenylated replication-dependent histone mRNA transcripts as well via reduced expression of the E2F-dependent histone-specific transcription regulator nuclear protein Ataxia Telangiectasia locus (NPAT). We could show that NPAT can directly interact with the positive transcription elongation factor b (P-TEFb) and thus recruits CDK9 to replication-dependent histone genes playing a role in proper replication-dependent histone mRNA 3’ end processing. These latest results uncover a mechanism of a regulated switch form processed replication-dependent histone mRNA transcripts to polyadenylated ones during a normal cellular process.

(15)

1. General Introduction

1.1 RNA Polymerase II-dependent transcription and mRNA processing

In eukaryotes protein coding genes are transcribed by RNA Polymerase II (RNAPII) from DNA into messenger RNA (mRNA) in the nucleus. Subsequently, the mRNA is transported into the cytoplasm and translated into proteins. The transcript generation by RNAPII takes place in a so-called transcription cycle that includes the three main phases initiation, elongation, and termination. Before the first step takes place, general transcription factors (GTFs) build a platform at the promoter of the respective gene and recruit RNAPII to form the preinitiation complex (PIC). In detail, the PIC includes RNAPII, the general factors TFIID, IIB, IIE, IIF, IIH and additional cofactors (Orphanides et al., 1996). TFIIH, the last initiation factor that binds to the PIC, has ATPase, helicase and kinase activity (Svejstrup et al., 1996). After formation of an open complex between RNAPII and the melted double-stranded DNA, which is an ATP-dependent process, the first phosphodiester bond is built and the presence of nucleotide triphosphates (NTPs) allows RNAPII to clear the promoter. The kinase activity of TFIIH finally leads to the phoshorylation of the carboxy-terminal domain (CTD) of the largest subunit of RNAPII (RBP1), at serine residue 5 (Ser5) being an important part of transcriptional regulation (Lee and Young, 2000). This event will be described in detail in section 1.1.3 on page 4ff. In the early stage of elongation the transcription elongation complex (TEC) is instable and has the tendency to release the transcribed mRNA, a process which is called abortive initiation (Sims III et al., 2004).

It was shown that after synthesis of a 23 nucleotide long transcript the slippage becomes undetectable and a stabilized TEC is formed (Pal and Luse, 2003). The promoter escape may be followed by an additional regulatory step, promoter-proximal pausing, which may be subsequently followed by productive elongation (Saunders et al., 2006). The transcription cycle finally ends with the termination phase in which the mRNA is cleaved and polyadenylated and RNAPII is removed from the gene (Proudfoot et al., 2002).

1.1.1 Regulation of transcription by promoter-proximal pausing

Promoter-proximal pausing, a rate limiting and regulatory step, occurs when the transcript reaches a size of 20-50 bases (Bentley, 2005). This phenomenon was first

(16)

demonstrated for the heat-shock gene hsp70 of Drosophila melanogaster (Gilmour and Lis, 1986). Before RNAPII commits into productive elongation, promoter-proximal pausing functions as a checkpoint and is controlled by specific negative elongation factors that will now be introduced (Saunders et al., 2006). The entry of RNAPII into productive elongation is repressed by two important negative factors the DRB sensitivity-inducing factor (DSIF) consisting of the subunits suppressor of Ty4 (Spt4p) in yeast or suppressor of Ty4 homolog (SUPT4H) in humans, Spt5p or SUPT5H respectively and the negative elongation factor (NELF) which comprises four subunits NELF-A, B, C/D and E (Narita et al., 2003; Wada et al., 1998a; Yamaguchi et al., 1999). DSIF can bind to the hypophosphorylated form of RNAPII (RNAPII_a) and NELF then binds to the preformed binary DSIF/RNAPII complex and the nascent RNA which leads to an increased pausing time and a decreased transcription rate of RNAPII (Renner et al., 2001; Yamaguchi et al., 2002).

It is postulated that the pausing of RNAPII mediated by DSIF and NELF is needed to allow the recruitment of the capping enzyme (CE) and hence the addition of a 7-methyl guanosine cap to the 5’ end of the nascent transcript (Sims III et al., 2004).

The CE associates with the Ser5-phosphorylated form of the CTD of RNAPII and with SUPT5H (Rodriguez et al., 2000; Wen and Shatkin, 1999). The efficient release of paused RNAPII into the productive elongation phase is largely controlled by the positive transcription elongation factor b (P-TEFb). P-TEFb phosphorylates the CTD of RNAPII at serine residue 2 (Ser2) as well as the human DSIF subunit SUPT5H.

SUPT5H is phosphorylated at a C-terminal repeat similar to the RNAPII-CTD and this modification leads to a conversion from a negative to a positive function of DSIF in transcriptional elongation (Peterlin and Price, 2006; Yamada et al., 2006). In addition, P-TEFb phosphorylates NELF-E, thereby releasing RNAPII from the repressive complex (Fujinaga et al., 2004; Peterlin and Price, 2006). Furthermore, the chromatin specific transcription elongation factor FACT (facilitates chromatin transcription) that was originally identified for its role in allowing elongation through chromatin, cooperates with P-TEFb to remove the DSIF/NELF-mediated negative regulation of transcription (Wada et al., 2000). The combined effects of different P-TEFb phosphorylation events relieves the repressing action of these negative elongation factors on RNAPII which allows the change from promoter-proximal pausing to productive elongation (Fujinaga et al., 2004; Ivanov et al., 2000; Kim and Sharp, 2001;

Renner et al., 2001; Wada et al., 1998b; Yamaguchi et al., 1999). P-TEFb thus has

(17)

separate functions to promote transcription. On the one hand it blocks the activity of repressors in the early stage of elongation and on the other hand it phosphorylates the CTD during productive elongation (Wood and Shilatifard, 2006). The fact that productive elongation requires a variety of additional elongation factors shows that this is a highly regulated process.

1.1.2 The positive transcription elongation factor b (P-TEFb)

The positive transcription elongation factor b was identified as a factor that can stimulate 5,6-dichloro-1-β-D-ribofuranosyl-benzimidazole (DRB) sensitive transcription and is required for RNAPII to synthesize long transcripts (Marshall and Price, 1992). The complex is a heterodimer composed of the cyclin-dependent kinase 9 (CDK9), a cell division cycle protein 2 (CDC2)-like Serine/Threonine kinase, and one of four cyclin partners, Cyclin T1, T2a, T2b or Cyclin K (Fu et al., 1999; Peng et al., 1998b; Peng et al., 1998a). CDK9 was identified in 1994 and first named PITALRE, based on a conserved sequence found in CDC2 and related kinases, due to its unknown function and interaction partners (Grana et al., 1994). In addition to the well known 42 kDa form of the CDK9 protein a second, larger form, CDK9(55), was identified that is translated from an mRNA that is generated by using a second upstream promoter of the CDK9 gene (Shore et al., 2003). As a general transcription factor P-TEFb is required for the efficient transcription of most genes. Its main function is the Ser2-phosphorylation of the CTD of RNAPII thereby promoting transcriptional elongation and other co- transcriptional processes (Marshall et al., 1996).

Compared to the functions of P-TEFb in mammals there are two homologs that combine its activities in Saccharomyces cerevisiae which include the Bur1p/Bur2p and the Ctk kinase complexes (consisting of Ctk1p, Ctk2p and Ctk3p) that may both function in the phosphorylation of the CTD (Wood and Shilatifard, 2006).

P-TEFb can be found either in an active or in an inactive complex. The core active complex comprises CDK9, Cyclin T1 or T2 and the bromodomain containing 4 protein (BRD4) (Jang et al., 2005). BRD4 was shown to recruit P-TEFb to transcriptional templates in vitro and in vivo and since BRD4 binds acetylated residues on histones it may facilitate P-TEFb recruitment to active genes (Bres et al., 2008; Wu and Chiang, 2007; Yang et al., 2005).

Besides the core active P-TEFb complex an inactive “large” P-TEFb complex exists and contains at least 50% of CDK9 in HeLa cells. In addition to CDK9 and

(18)

Cyclin T1 or T2 the large complex comprises the 7SK small nuclear RNA (7SK snRNA) and the hexamethylene bisacetamide-induced protein 1 or 2 (HEXIM1,2) (Michels et al., 2003; Michels et al., 2004; Nguyen et al., 2001; Yang et al., 2001; Yik et al., 2003). It is postulated that a single molecule of 7SK snRNA binds a dimer of HEXIM1 or HEXIM2 thus changing its conformation so that it exposes a binding domain for P-TEFb. Phosphorylation of T186 of the T-loop of CDK9 allows the binding of two complexes to the HEXIM proteins resulting in an inactive P-TEFb complex (Li et al., 2005b; Marshall and Grana, 2006). Another component of the inactive P-TEFb complex is La ribonucleoprotein domain family, member 7 (LARP7) also designated as PIP7S (P-TEFb interaction protein for 7SK stability) (He et al., 2008;

Krueger et al., 2008; Markert et al., 2008). As a stable component of the 7SK snRNP it stays bound to 7SK snRNP when P-TEFb is activated and released and its expression is essential for maintaining 7SK stability and the integrity of the large complex (Krueger et al., 2008).

A further possibility of regulating P-TEFb activity is the acetylation of CDK9 with the major acetylation site being at lysine 44. This posttranslational modification was shown to enhance the ability of P-TEFb to phosphorylate the serine 2 residue of the CTD of RNAPII (Fu et al., 2007).

1.1.3. The “CTD-code” of RNAPII

In addition to its catalytical core, the largest subunit of eukaryotic RNAPII (RBP1) contains an unusual carboxy-terminal domain (CTD) that comprises 52 repeated heptapeptides in humans and mice with the consensus sequence Tyr-Ser-Pro- Thr-Ser-Pro-Ser (Y₁S₂P₃T₄S₅P₆S₇) (Corden et al., 1985; Corden, 1990). This sequence is evolutionary conserved but the number of repeats varies from 26 in yeast, 32 in nematodes, 45 in Drosophila melanogaster and 52 in mammals (Egloff and Murphy, 2008). It is a target for phosphorylation which can act as a platform for proteins that are implicated in different co-transcriptional processes (Brookes and Pombo, 2009;

Dahmus, 1995).

Two forms of RNAPII either containing a hypo- or a hyperphosphorylated CTD (RNAPII_a and RNAPII₀ respectively) exist (Dahmus, 1981). The non- or hypophosphorylated RNAPII is associated with the PIC at the promoter whereas the hyperphosphorylated form is part of the elongating complex. As already mentioned in section 1.1 on page 1 and section 1.1.2 on page 3f. phosphorylation mainly occurs on

(19)

Ser2 and Ser5 of the heptapeptide but was recently found on Ser7 (Chapman et al., 2007; Egloff et al., 2007). Phosphorylation of Thr4 and Tyr1 have also been postulated, as has glycosylation of Thr4. Ser5 phosphorylation occurs near the 5’ end of genes and is carried out by the cyclin-dependent kinase 7 (CDK7) a component of the general transcription factor TFHII that binds to the hypophosphorylated form of RNAPII (Komarnitsky et al., 2000). This modification allows the binding of the guanyltransferase that is responsible for addition of the 7-methyl guanosine cap to the 5’

end of the newly synthesized mRNA (Cho et al., 1998). When promoter-proximal pausing occurs after promoter escape CDK9 as part of P-TEFb phosphorylates the CTD at Ser2 allowing RNAPII for productive elongation (Marshall et al., 1996). 3’ end processing factors like the protein 1 of cleavage and polyadenylation factor I (PCF11) were shown to directly bind to P-Ser2 repeats containing CTDs (Meinhart and Cramer, 2004). Chromatin Immunoprecipitation (ChIP) results showed that P-Ser5 of the CTD predominates near the transcription start site of a gene, whereas in the middle of a gene both Ser2 and Ser5 are phosphorylated and near the 3’ end Ser2 is extensively phosphorylated (Phatnani and Greenleaf, 2006). Phosphorylation of Ser7 appears to play a gene-specific role being required for the proper expression of snRNA genes but not protein-coding genes. Furthermore, it seems to be critical for recruiting the Integrator complex which is necessary for the 3’ end processing of snRNAs (Egloff et al., 2007). Recent findings indicated that the kinase homolog to CDC28 (Kin28p) in yeast and its homolog CDK7 in humans are the kinases responsible for phosphorylation of Ser7 (Akhtar et al., 2009; Boeing et al., 2010).

Since the phosphorylation of the CTD is a dynamic process, also dephosphorylation must be performed by specific phosphatases. The TFIIF-associated CTD phosphatase 1 (FCP1) and the small CTD phosphatase 1 (SCP1) were reported to dephosphorylate Ser2 and Ser5 respectively of the heptapeptide sequence (Egloff and Murphy, 2008; Meinhart et al., 2005; Yeo et al., 2003; Yeo et al., 2005).

Phosphorylated RNAPII is associated with pre-mRNA processing factors like the 5’ end capping enzyme, splicing factors, cleavage and polyadenylation factors (Bentley, 2005). Up to now, 8 different phosphorylation patterns within the repeat are possible (unphosphorylated CTD, combinations of P-Ser2, P-Ser5 and P-Ser7 CTD).

These different modification patterns appear to play a role in the recruitment of specific regulatory factors that control the activity of transcribing RNAPII (Corden, 2007).

Thus, differential posttranslational modifications of the CTD residues within the

(20)

heptapeptide sequence generate a so-called “CTD code” that serves as a platform for a variety of proteins that are timely recruited during the transcription cycle (Egloff and Murphy, 2008). In addition to the phosphorylation pattern the cis/trans isomerization of the proline residues Pro3 and Pro6 by the peptidyl-prolyl cis/trans isomerase, NIMA- interacting 1 protein (Pin1p) which binds to Ser2- and Ser5-phosphorylated CTD and change the structure of the CTD likely also add to the complexity of the “CTD code”

(Buratowski, 2003).

Each of these different possible conformation patterns of the CTD is thought to play its own role in the regulation of RNAPII-dependent transcription by being responsible for specific transcriptional steps and co-transcriptional pre-mRNA processing events. Theoretically, every residue is a potential target for some kind of posttranslational modification. Thus, multiple CTD modifications help to recruit specific proteins which play important roles in the regulation of transcription (illustrated in Fig 1 on page 6).

Fig 1: RNAPII-dependent transcription and the CTD phosphorylation patterns. When the PIC is built RNAPII is present in a hypophosphorylated form. TFIIH then phosphorylates Ser5 and in some genes Ser7 which leads to promoter clearance. Pausing takes place upon binding of the negative elongation factors NELF and DSIF, the capping enzyme (CE) is recruited and 5’ capping takes place.

Upon recruitment of P-TEFb, Ser2, DSIF and NELF are phosphorylated, leading to the dissociation of NELF, the conversion of DSIF into a positive transcription elongation factor and consequently to productive elongation (modified from Peterlin and Price, 2006; Phatnani and Greenleaf, 2006).

1.1.4 Co-transcriptional mRNA processing

Several pre-mRNA processing events including 5’ end capping, splicing and 3’

end formation by cleavage and polyadenylation occur co-transcriptionally (Bentley, 2005; Sisodia et al., 1987). The CTD of RNAPII provides a landing pad for mRNA processing factors and plays a central role in the coupling of processing and

(21)

transcription since the truncation of the CTD was shown to result in defects in capping, cleavage/polyadenylation and splicing (McCracken et al., 1997).

The first event of co-transcriptional processing during the transcription cycle takes place at the step of transcriptional initiation and requires Ser5-phosphorylated CTD. As already mentioned in section 1.1.3 on page 4ff. this modification is recognized by the capping machinery implicating the three different enzymes RNA triphosphatase, guanylyltransferase and 7-methyltransferase in yeast or rather two enzymes, the capping enzyme, that implicates triphosphatase and guanylyltransferase activities, and the 7- methyltransferase in mammalian (Ho and Shuman, 1999; Schroeder et al., 2000;

Shuman, 2001). It was shown that the human capping enzyme can stimulate promoter escape of RNAPII by relieving transcriptional repression by NELF and that its recruitment is enhanced by directly binding to SUPT5H (Mandal et al., 2004). Since human capping enzymes were found at 5’ ends as well as throughout genes and even in 3’ regions downstream of the poly(A) site, they seem to stay bound to the RNAPII platform after capping has taken place and may therefore even influence elongation, termination and 3’ end processing (Glover-Cutter et al., 2008; Perales and Bentley, 2009).

Splicing factors are recruited to the nascent transcript as well and either remove most of the introns co-transcriptionally or mark others for post-transcriptional splicing (Wetterberg et al., 2001). For splicing U1 snRNP first binds to the GU sequence at the 5’ splice site of the intron and U2 small nuclear RNA auxiliary factor (U2AF) to the AG 3’ splice site. U2 snRNP then base pairs with the branch site, the triple snRNP U4- U6/U5 is recruited and rearranged to assume the catalytically active conformation as U1 and U4 are discarded (Perales and Bentley, 2009). U1 snRNP and other splicing regulatory proteins were shown to be recruited to genes and to interact with the CTD of RNAPII resulting in efficient splicing (Das et al., 2007; Listerman et al., 2006).

Additionally, the knockdown of the chromodomain protein 1 (CHD1) and the decrease in H3K4me3 levels leads to a diminished association of U2 snRNP with chromatin and results in less efficient pre-mRNA splicing in vivo showing that H3K4me3 is crucial for efficient recruitment of spliceosomal proteins via the CHD1 protein (Sims III et al., 2007).

Cleavage and polyadenylation of the transcript also occurs co-transcriptionally and is performed by a large cleavage/poly(A) complex. In yeast this complex is recruited to elongating RNAPII in part by binding of the components polyadenylation

(22)

and cleavage factor subunit 11 (Pcf11p) which is a subunit of the cleavage factor 1A (CF1A) and the termination factor regulator of Ty1 transposition 103 (Rtt103p) to the Ser2-phosphorylated CTD (Kim et al., 2004; Licatalosi et al., 2002; Meinhart and Cramer, 2004; Proudfoot, 2004). In mammals the cleavage factor CPSF73 was found at both the 5’ and 3’ ends of genes and is thought to first bind with TFIID to the promoter and is then handed off to RNAPII and travels with it (Glover-Cutter et al., 2008).

Recently, purification of the human 3’ end processing complex identified approximately 85 proteins including the polymerase associated factor (PAF) complex subunit cell division cycle 73 (CDC73) and the FACT subunit suppressor of Ty16 homolog (SUPT16H) implicating a connection between 3’ end processing and transcriptional elongation in human cells (Shi et al., 2009). The cleavage and 3’ end formation events in mammals will be explained in detail in the next sections.

1.1.4.1 Mammalian 3’ end processing of polyadenylated mRNAs

In higher eukaryotes the canonical polyadenylation signal AAUAAA is recognized and bound by the cleavage and polyadenylation specificity factor (CPSF) that consists of at least five subunits (CPSF160, CPSF100, CPSF73, CPSF30 and FIP1).

Another canonical sequence, the downstream sequence element (DSE) that is rich in G/U or U residues is located up to 30 nt downstream of the cleavage site and is bound by the cleavage stimulating factor (CSTF). After assembly of the basal 3’ end processing machinery that consists of the cleavage factors I and II (CFI and CFII), CPSF73 catalyses the endonucleolytic cleavage at the cleavage site, 10-30 nt downstream of the poly(A) signal (Danckwardt et al., 2008; Shi et al., 2009; Weiner, 2005). Afterwards, the poly(A) polymerase (PAP) adds approximately 250 A- nucleotides to the 3’ end of the cleaved transcript (Danckwardt et al., 2008). Additional upstream or downstream signals can facilitate the cleavage that often occurs 3’ to the dinucleotide CA (Weiner, 2005). CPSF was shown to directly interact with the transcription factor TFIID already being associated with transcription in the preinitiation complex (Dantonel et al., 1997).

In contrast to all other known eukaryotic mRNAs, replication-dependent histone mRNAs are not polyadenylated and their processing requires a different set of proteins.

The details of histone mRNA 3’ end processing and the functions of important factors being involved will be explained in section 1.3.2 on page 23ff.

(23)

1.1.4.2 The importance of 3’ end processing

Eukaryotic pre-mRNA 3’ end processing is a very important process that has several functions crucial for cell growth and viability. Polyadenylation of the 3’ end of the pre-mRNA is important for its transport from the nucleus into the cytoplasm (Vinciguerra and Stutz, 2004). Furthermore, it functions in promoting stability of the mRNA by preventing degradation by 3’ exonucleases which is a principle mechanism controlling mRNA stability (Wickens et al., 1997). Both, the addition of a poly(A) tail and the binding of the poly(A)-binding protein (PABP) has been shown to prevent degradation of mRNAs in the cytoplasm (Ford et al., 1997). The translation of mRNAs into proteins is enhanced by the interaction of the poly(A) tail and the PABP with the 5’

cap (Wilusz et al., 2001). This interaction was shown to optimize the efficiency of translation in yeast (Preiss and Hentze, 1998). The 3’ end processing machinery interacts with transcription factors of the preinitiation complex and with the CTD of RNAPII thus potentially playing a role in controlling transcriptional initiation. In addition, the poly(A) signal is crucial for proper transcriptional termination (Bentley, 2002; Proudfoot, 2004). Conversely, defects in 3’ end processing can negatively influence transcription (Manley, 2002).

1.2 Chromatin structure and histone modifications

The genomic DNA of all eukaryotic cells is highly condensed and wrapped around core histone proteins to produce a condensed structure called the chromatin (Kornberg, 1974). One nucleosome, which is the fundamental unit of the chromatin structure, comprises a core of eight histone proteins: two each of histone 2A (H2A) and 2B (H2B), histone 3 (H3) and histone 4 (H4) whereby two H2A-H2B dimers interact with an H3-H4 tetramer. Around these nucleosomes 147 base pairs of DNA are twined (Kornberg, 1977; Luger et al., 1997). The H1 histone serves as a linker that associates with DNA between single nucleosomes which leads to a high folded 30 nm fiber (Allan et al., 1980; Thoma et al., 1979). Histones are small basic proteins that consist of a flexible positively charged N-terminus the so-called “histone tail” and a globular COOH-terminal domain that make up the nucleosome scaffold (Fischle et al., 2003;

Jenuwein and Allis, 2001). The N-termini protrude away from the DNA and are thus exposed on the nucleosome surface (Luger et al., 1997). The majority of chromatin is found in a condensed, compact form called heterochromatin that contains only a few

(24)

genes and is generally transcriptionally inactive. In comparison, the uncondensed regions of the genome that contain a high density of genes is referred to as euchromatin and generally more transcriptionally active (Henikoff, 2000; Richards and Elgin, 2002).

Today it is known that not only the structure of chromatin but also posttranslational modifications of the histone tails are linked to transcriptional regulation.

1.2.1 Histone modifications and the “histone code”

Given their position of the surface of the nucleosomes the histone tails function as acceptors for a variety of enzyme-catalyzed posttranslational modifications taking place at specific amino acid side chains including acetylation, methylation, ubiquitination and sumoylation of lysine residues, phosphorylation of serine and threonine residues, and methylation of arginine residues (Fischle et al., 2003; Turner, 2007) (illustrated in Fig 2 on page 11). There are two possible mechanisms how posttranslational modifications of histones can regulate transcriptional activity. The first model predicts that the different chromatin modifications result in changes in electrostatic charge i.e. acetylation that neutralizes positive charge or phosphorylation that adds a negative charge or from modified interactions between nucleosomes (Strahl and Allis, 2000). This change would then affect the ability of nucleosomes to bind to DNA and thereby alter the accessibility for DNA binding proteins such as transcription factors. The second and probably most important model is the change of the nucleosome surface by the attached chemical poststranslational modifications that serves as a platform for specific chromatin-binding proteins (Berger, 2007). Two classes of domains that interact with specific modified residues are the bromodomain which binds to acetylated lysine residues and the chromodomain that interacts with methylated lysine residues (Berger, 2002). The hypothesis that a network of posttranslational modifications including the number, type, combination and localization within the genome controls the status of a gene thereby controlling the biological outcome is termed the “histone code” (Berger, 2002; Jenuwein and Allis, 2001; Strahl and Allis, 2000). In addition, histone modifications can influence each other in a synergistic or antagonistic way leading to a complex and diverse imprint pattern (Jenuwein and Allis, 2001). There are histone modifications that are linked to transcriptional activation and others for transcriptional repression. The most studied example, histone actetylation, is balanced by histone acetyltransferases (HATs) and histone deacetylases (HDACs) and is generally associated with the promoters of actively transcribed genes (Li et al., 2007).

(25)

In the next section some specific other posttranslational modifications are pointed out and described in detail.

Fig 2: Posttranslational modifications of the core histones and their positioning. The histone tails can be methylated at lysines and arginines (green pentagons), phosphorylated at serines or threonines (yellow circles), acetylated at lysines (red triangles), or ubiquitinated at lysines (blue stars) (modified from Peterson and Laniel, 2004).

1.2.1.1 Histone H2B monoubiquitination (H2Bub1)

Monoubiquitination of histone H2B (H2Bub1) was shown to be associated with transcriptionally active DNA and appears to occur primarily co-transcriptionally (Davie and Murphy, 1990; Osley, 2006). In humans H2Bub1 is preferentially found in the transcribed regions of highly expressed genes and was reported in yeast to be required for efficient reassembly of nucleosomes during elongation. Therefore H2Bub1 was suggested to be linked to transcriptional elongation in both yeast and humans (Fleming et al., 2008; Minsky et al., 2008).

The ubiquitination process requires three separate enzymatic activities (Hochstrasser, 1996). First, ubiquitin, a 76 amino acid protein, is activated by an ubiquitin activating enzyme (E1) in an ATP-dependent manner. Secondly, it is conjugated to a cysteine residue of a ubiquitin conjugating enzyme (E2) via a thioester bond and finally transferred to a target lysine residue by a ubiquitin-protein-isopeptide ligase (E3) (Weake and Workman, 2008). Whereas polyubiquitination generally targets a protein for degradation via the 26S proteasome, monoubiquitination is generally associated with a change in protein function (Pickart, 2001).

(26)

In yeast (Saccharomyces cerevisiae) H2Bub1 takes place at lysine residue 123 (K123) whereas in humans it occurs at lysine residue 120 (K120) (Osley, 2006). Less than 10% of the total histone H2B are monoubiquitinated and this modification is rapidly removed by ubiquitin proteases (UBPs) and is thus a highly dynamic process (Osley, 2004; Zhang, 2003). In yeast, Ubp8p, a component of the Spt-Ada-Gcn5- acetyltransferase (SAGA) complex deubiquitinates H2B in vitro and in vivo and in humans ubiquitin specific protease 22 (USP22), a subunit of the human SAGA complex was shown to be the ortholog of yeast Ubp8p (Henry et al., 2003; Zhang et al., 2008).

Interestingly, both ubiquitination and deubiquitination of histone H2B appear to be involved in transcriptional activation (Henry et al., 2003).

In yeast, the E2 enzyme for H2Bub1 in vivo was shown to be radiation-sensitive mutant protein 6 (Rad6p) and the interacting E3 ligase was identified to be a RING finger protein called brefeldin A sensitive protein 1 (Bre1p) (Hwang et al., 2003;

Robzyk et al., 2000; Wood et al., 2003a). Based on sequence alignment two homologs for the yeast Rad6p protein have been found in humans, namely hHR6A and hHR6B or also called UBE2A and UBE2B according to the new nomenclature (Koken et al., 1991). Up to now it is debated if these are the E2 enzymes responsible for H2Bub1 in vivo or if rather another ubiqutin conjugase, UBCH6, could be implicated in this process (Pavri et al., 2006; Zhu et al., 2005b). Concerning the E3 ligase two RING finger proteins, RNF20 and RNF40, that share sequence homology with yeast Bre1p were shown to both be essential for monoubiquitination of histone H2B (Kim et al., 2009;

Zhu et al., 2005b).

Rad6p can associate with the elongating form of RNAPII and this interaction is dependent upon Bre1p and the Paf1p complex in yeast (Xiao et al., 2005). The yeast Paf1p complex, consisting of five different proteins (Paf1p, Rtf1p, Ctr9p, Cdc73p and Leo1p) was identified to bind to Ser5-P CTD of RNAPII and to interact with Bre1p to recruit Rad6p to the promoter region of actively transcribed genes (Laribee et al., 2007).

Rad6p is also phosphorylated and thereby activated by the Bur1p/Bur2p kinase complex which in this way is implicated in controlling monoubiquitination of H2B (Wood et al., 2005; Wood and Shilatifard, 2006). These observations show that H2Bub1 is a histone mark that is coupled to RNAPII transcriptional elongation in both yeast and humans.

(27)

1.2.1.2 Histone H3 lysine 4 trimethylation (H3K4me3)

Methylation of histones occurs either on arginine and/or lysine residues, whereas arginine methylation of histones is involved in gene activation but lysine methylation instead can influence either activation or repression (Berger, 2002). The lysine residues of histone H3 or H4 are methylated by lysine methylating transferases (KMTs) and can be mono- di- or trimethylated. Each level of modification can have a different biological outcome showing the complexity of this posttranslational modification (Berger, 2007).

Trimethylation of histone H3 (H3K4me3) is associated with highly transcriptionally active genes whereas dimethylation of histone H3 (H3K4me2) occurs at both inactive and active genes (Berger, 2007; Bernstein et al., 2002; Santos-Rosa et al., 2002). Concerning the localization of this posttranslational modification, it was shown that H3K4me3 occurs near the transcription start site (TSS) and the 5’ end of the open reading frame (ORF), H3K4me2 is found throughout the whole body of the gene but peaks in the middle and H3K4me1 is enriched at the 3’ end of genes (Li et al., 2007;

Pokholok et al., 2005). Mono-, di- and trimethylation of histone H3 is performed by the complex of proteins associated with Set1p (COMPASS) in yeast, a homolog of the human methyltransferases SET1 and the mixed lineage leukemia (MLL) protein family (Schneider et al., 2005). Human MLL complexes contain the subunits absent, small, or homeotic-like protein (ASH2L), retinoblastoma binding protein 5 (RBBP5) and WD repeat domain 5 protein (WDR5) and knockdown of ASH2L results in a global reduction of H3K4me3 (Steward et al., 2006). The MLL1 gene is involved in numerous translocations that are found in several human acute leukemias (Rowley, 1998; Tenney and Shilatifard, 2005). The fusion proteins of such translocated MLL genes loose their H3K4 methyltransferase domain but are still able to bind to homeobox (HOX) genes and are associated with an increase in their expression and thus dysregulation (Guenther et al., 2005; Krivtsov et al., 2008; Rozovskaia et al., 2001).

Since H3K4 methylation does not affect transcription elongation in vitro per se and in addition the methyltransferase Set1p does not affect the processivity of RNAPII in yeast in vivo, the precise function of H3K4 methylation in transcription regulation is not yet known (Mason and Struhl, 2003; Pavri et al., 2006). The importance of H3K4 methylation thus seems to lie primarily in a signalling function (Li et al., 2007). This is supported by the fact that chromatin-remodeling factors and histone modification complexes that contain plant homeodomain (PHD) fingers can directly recognize and

(28)

bind to H3K4me3 (Zhang, 2006) and recruit complexes to activate or repress transcription.

The methylation status of H3K4 plays a role in either recruiting, activating or repressing effectors. H3K4me2 and H3K4me3 both recruit acetyltransferases directly or via binding of Chd1p, an elongation-related chromatin-remodeling factor, and the methyltransferase Set1p as part of COMPASS which leads to an increase in other histone acetylations and trimethylated H3K4 respectively (Berger, 2007; Pray-Grant et al., 2005; Sims III and Reinberg, 2006). In humans CHD1 associated with H3K4me3 assists in recruiting pre-spliceosomal components like U2 snRNP to chromatin resulting in facilitating pre-mRNA maturation (Sims III et al., 2007; Sims III and Reinberg, 2009). Importantly, via this posttranslational modification the rate but not the catalysis of splicing in human cells is affected. Thus, the role of H3K4me3 seems to lie mainly in serving as a mark for proteins that are implemented in the recruitment of complexes that influence transcription or pre-mRNA maturation.

In contrast, only H3K4me3 serves as a signal to recruit repressing effectors like deacetylases and demethylases leading to repression of gene activity (Berger, 2007; Shi et al., 2006; Huang et al., 2006). Like this, the posttranslational modification of one residue on a histone can have multiple biologic outcomes.

1.2.1.3 Histone H3 lysine 36 trimethylation (H3K36me3)

Di- and trimethylation of lysine 36 in histone H3 (H3K36me2 and H3K36me3) are both coupled to transcriptional elongation.

In Saccharomyces cerevisiae the histone methyltransferase Set2p was found to be the specific enzyme for methylating histone H3 on lysine residue 36 (di- and trimethylation) (Strahl et al., 2002). It is recruited to the CTD phosphorylated by the Ctk complex (Wood and Shilatifard, 2006). Set2p was shown to directly interact with elongating RNAPII by binding to dual P-Ser2/P-Ser5 phosphorylated CTD but not to unphosphorylated RNAPII in vitro (Li et al., 2003; Xiao et al., 2003). In addition, ChIP analysis revealed the association of Set2p and H3K36 methylation with the coding region of genes and the correlation with active transcription (Krogan et al., 2003).

In Drosophila melanogaster H3K36me2 peaks near the promoter region of genes whereas H3K36me3 is accumulated towards the 3’ end of genes. Interestingly, these methylation events are performed by two differents KMTs, namely maternal effect sterile family member 4 (dMes-4) and huntingtin interacting protein b (dHypb)

(29)

respectively (Bell et al., 2007). In humans the methyltransferases HYPB, SET and MYND domain containing 2 (SMYD2) and nuclear receptor-binding SET domain- containing protein 1 (NSD1) appear to take over these roles since they can methylate H3K36 in vitro (Brown et al., 2006; Edmunds et al., 2008; Rayasam et al., 2003; Sun et al., 2005). The enzyme that is relevant in vivo and if it catalyses mono-, di- or trimethylation of H3K36 still remains to be elucidated.

Concordantly, these studies show that in both yeast and mammals, P-Ser2 of the CTD is necessary to recruit the histone methyltransferase responsible for the trimethylation of H3K36 (Krogan et al., 2003; Li et al., 2005a).

The role of H3K36me3 during transcriptional elongation was shown in yeast to recruit histone deacetylases specifically to the transcribed region in order to suppress improper internal initiation (Carrozza et al., 2005). Otherwise transcription initiation on cryptic promoters in the body of genes would lead to 5’ truncated transcripts and thus aberrant protein production.

1.2.1.4 Histone modification crosstalk

Histone modifications can influence one another. For example, one modification can function to recruit an enzyme or activate an enzymatic activity that is necessary to generate another histone modification (Suganuma and Workman, 2008). This so-called trans-histone effect is well known for H2Bub1 in being a prerequisite for H3K4me3 and H3K79me3 and is highly conserved from yeast to human (Dover et al., 2002; Fischle et al., 2003; Ng et al., 2003; Sun and Allis, 2002; Wood et al., 2005). Rad6p was shown to be essential for H3K4me3 through H2Bub1 on lysine 123 in yeast in a unidirectional regulatory pathway (Sun and Allis, 2002). Not only H3K4me3 by the Set1p-containing COMPASS complex but also H3K79me3 by a methyltransferase called disruptor of telomeric silencing 1 (Dot1p) was shown to be dependent upon H2Bub1 (Briggs et al., 2002; Ng et al., 2002; Singer et al., 1998). Interestingly, H2Bub1 specifically affects di- and trimethylation of H3K4 and H3K79 but not monomethylation in yeast. Thus, it is thought that not the enzyme recruitment is affected but in fact the processivity of the histone metyltransferase (Dehe et al., 2005; Shahbazian et al., 2005). By using chemically ubiquitinated H2B, McGinty et al. recently showed that H2Bub1 directly activates DOT1L-mediated intranucleosomal methylation of H3K79 in humans and that this posttranslational modification is carried out by the catalytic domain of DOT1L (McGinty et al., 2008). However, the recruitment of DOT1L was not dependent upon

Epigenetic Regulation of Replication-Dependent Histone mRNA 3 End Processing